PRV seat melted

[prq123]

We had a pilot PRV stick full open for about 12 hours. When the PRV was taken offline fore repairs it was noticed that the PRV soft seat had partially melted along the outside edge. The melting temperature of the soft seat is > 160 deg C, however, the measured gas temperature during the relief (and normal operating temperature) was only 100 deg C. The gas temperature has never exceed 110 deg C during the entire PRV life. The PRV discharges to a flare header and the fluid is clean natural gas. The upstream pressure when it was stuck open was around 67 bar.

Does anyone have ideas on what could have caused the PRV seat to melt? Could it have melted when the PRV was stuck open?

 

[Latexman]

Instead of melting, maybe it dissolved by some chemical attack. It looks about the same. Is the soft seat compatible with the process?

Good luck,
Latexman

 

[prq123]

An update: The damaged soft seats were sent to a lab for detailed analysis. The lab confirmed the melting was due to heat damage (temperature > 200 C). However, there are 4 independent temperature indicators on this stream and none have indicated temperatures above 110 C. The PRVs are on the discharge line of a centrifugal compressor. Adjacent PRVs (in the same service) were removed and examined, and found no signs of melting.

Any ideas on what caused the damage? (The gas is a clean gas, only N2, methane, ethane).

 

[786392]

Dear
Just a wild thought (if I read your latest post info correctly)

Only temperature indicators are mentioned not recorders and the point location temperature may rise if there is

Centrifugal compressor running continuously at no flow condition,without intervention.

Consequent high temperature may result via kinetic energy accumulation;momentarily reaching to capability of melting PRV seat& make it pass.

Best Regards
Qalander(Chem)

 

[prq123]

We have 4 DCS temperature indicators/recorders at the immediate discharge of the compressor (before PRV takeoff) so we know the temperature history throughout the life of the PRV. PRV has been in service for less than 1 year and no temperataure above 110 C have been recorded.

 

[SeanB]

Is your PSV steam or electrically traced by any chance?

 

[prq123]

No, it is not steam/electrically traced.

 

[786392]

Dear then most presumably

either your sample's analysis indicting melting of seat seems incorrect(look for what 'latex' said in his last post)

or the DCS system temperature indicator cum recorder could not catch the very short duration event!

Best Regards
Qalander(Chem)

 

[prq123]

FYI - the sample was sent to two independent labs. Both labs concluded the damage was due to high temperatures. The DCS historical data archives data in 5 second scans, so it is very unlikely that a high temperature reading would not be picked up.

 

[786392]

Dear
I have only two of the possibilities left with(and it should come out from other learned forum colleagues if any more credible scenario could be envisaged)

Since during '12 hour sticking span'it makes 8640 readings recorded for each of the temperature indicator/recorder i.e. for 4 (4*8640=)34560 readings; all checked thoroughly to be between 100~110 Celsius and below 160 Celsius as this seems very difficult to me.(most probably no sensing piont at or in closed vicinity of this pilot PRV seat)

1)High Velocity initial erosion may resemble melting or momentarily might create high localized temperature only due to friction at the seat(Do Check in case any fine powdery black substance exists at the inner walls of some natural gas pipelines also accompanied somehow dislodged if existed unnoticeable!)

2)While stuck in open, any high temperature material could have been sucked back in from other pipelines connected or or even flame front from flare stack if no flash back/flame back arresting system(sealing arrangement) was in place.


Best Regards
Qalander(Chem)

 

[786392]

If the analytical approach forwarded was correct/accurate please intimate or otherwise?

Best Regards
Qalander(Chem)

 

[prq123]

To address your previous query:
1) The DCS readings can be exported in Excel numerical format. It is easy to find the maximum values of the readings in Excel.
2) We had an identical PRV (in a duplicate unit) also lift for only a few minutes about 1 month later. When this PRV was inspected it had the same heat damange. The gas has been through several filters (down to 10 micron level) before it get to the PRV so is clean gas. The unit is part of gas treating plant.
3) We have checked other PRVs connected to the same flare header (which have not lifted) and found no damage. We have DCS temperature recorders at the base of the flare stack and have recorded no high temperature in the flare header.

So far the reason for the damage is still a mystery and is under investigation.

 

[moltenmetal]

Perhaps your DCS temperature indications are made using an RTD or thermocouple so poorly coupled to the flowing gas that a gas temperature of 200+ C only gives a reading of 110 C?!

I've seen plenty of gas temperature measurements made incorrectly. It would not suprise me a bit if this system had some huge thick-walled thermowells which were so thermally coupled to the piping that the TE would have difficulty in reading any fluid temperature significantly different than that of the pipe! This would make them likely to miss transient high temperature events. Thermocouples not bottomed out in their wells, inserted dry without heat transfer grease etc., would tend to make a bad situation even worse.

 

[786392]

Thanks indeed for this important info update,

Kindly do read my(16th Jan post's)Closing sentence(within brackets);indicating that your sensing is not at both PRV's melted seat vicinity to authenticate the actual point temperature of meting seats area.

Thus the problem solving/ decision making process may not go forward in a nice& steady way to troubleshoot the issue.

Please appreciate.

Best Regards
Qalander(Chem)

 

[prq123]

1) Molten-As mentioned previously, the PRV's are at the discharge of a centrifugal compressor. Based on the acutal measured power absorbed by compressor it is impossbile for the temperature to exceed 160 C. The measured temperature agrees with Hysys simulations of the compressors based on the compressor efficiency/head.
2) Qalander - The temperature measuresment is at the immediate discharge of the compressor, before the PRV take off line. This should be the hottest gas (at discharge of compressor). If you can come up with a thermodynamic basis as to why the gas at the PRV inlet should be higher than the compressor discharge, I would like to hear it.

 

[786392]

If I've not mistaken(mis-read)
your previous post 27th Jan 4:56

item 3) 2nd sentence indicated these DCS temperature recorders to be at bottom of flare stack,I believe

or I grossly mis-read/mis-understood the same!

Best Regards
Qalander(Chem)

 

[prq123]

My previous statement on the flare header DCS temperature measurements was to rule out any high temperatures coming from the flare header (as you suggested as possible cause earlier).

We have separate DCS temperature measurement at the discharge of the compressor.

 

[786392]

Dear
Is there any record of
(the Two) PRV's suffering melted seat damage(s)upstream side gas filter element(s) differential pressure Monitoring & Any sudden sharp downward differential pressure variations across any one of these gas filter element(s)?

If such is the case that either no clear recording exists or frequent repairs/replacement record is known

then sudden erosion with high velocity particles cannot be ruled out.(and which may exhibit melting like appearance!)
Hope this helps

Best Regards
Qalander(Chem)

 

[PAAB]

What is the material of the 'soft' seat?

 

[prq123]

OK people, thanks for your help. Please no more questions.